Feature Request / Improvement
Feature request
Iceberg exposes many useful metadata tables, such as files, data_files, delete_files, manifests, snapshots, history, metadata_log_entries, refs, and others.
I would like to propose adding a new metadata table that lists the metadata tables available for a given Iceberg table.
Motivation / use case
Metadata tables are useful for debugging, observability, table maintenance, and understanding Iceberg table internals. For example, users can inspect files, manifests, snapshots, refs, and metadata log entries directly through SQL.
However, users currently need to know these metadata table names ahead of time from documentation, examples, or source code. It would be helpful to make metadata tables self-discoverable from SQL.
A self-describing metadata table would make Iceberg metadata easier to explore without adding engine-specific SQL syntax such as:
SHOW METADATA TABLES FOR catalog.db.table;
Instead, this would reuse Iceberg's existing metadata table model.
For example, in Spark:
SELECT * FROM catalog.db.table.metadata_tables;
Excepted output:
entries
files
data_files
delete_files
history
metadata_log_entries
snapshots
refs
manifests
partitions
all_data_files
all_delete_files
all_files
all_manifests
all_entries
position_deletes
metadata_tables
Query engine
Spark
Willingness to contribute
Feature Request / Improvement
Feature request
Iceberg exposes many useful metadata tables, such as
files,data_files,delete_files,manifests,snapshots,history,metadata_log_entries,refs, and others.I would like to propose adding a new metadata table that lists the metadata tables available for a given Iceberg table.
Motivation / use case
Metadata tables are useful for debugging, observability, table maintenance, and understanding Iceberg table internals. For example, users can inspect files, manifests, snapshots, refs, and metadata log entries directly through SQL.
However, users currently need to know these metadata table names ahead of time from documentation, examples, or source code. It would be helpful to make metadata tables self-discoverable from SQL.
A self-describing metadata table would make Iceberg metadata easier to explore without adding engine-specific SQL syntax such as:
Instead, this would reuse Iceberg's existing metadata table model.
For example, in Spark:
Excepted output:
Query engine
Spark
Willingness to contribute